Introduction: The hidden tax of cloud waste
Cloud adoption is no longer optional—it’s the default. According to the Flexera 2025 State of the Cloud Report, more than half of all enterprise and SMB workloads now run in the public cloud. Cloud has become the backbone of digital transformation, powering everything from e-commerce platforms and AI-driven analytics to global collaboration tools.
But with this adoption comes a silent tax: waste. For years, analysts have pointed out that a significant portion of cloud spend is unused, underutilized, or misallocated. In 2025, despite stronger FinOps adoption and more mature governance frameworks, organizations are still losing vast amounts of money. Flexera’s report estimates that 27% of IaaS and PaaS spend is wasted—equivalent to hundreds of billions globally]. Add in 24% wasted spend on SaaS licenses, and the picture gets even worse.
This waste is not just a financial issue. It slows innovation, skews forecasting, frustrates CFOs, and raises carbon footprints unnecessarily. Flexera provides a valuable high-level snapshot, but the reality on the ground is more complex. Waste is rarely uniform across workloads, regions, or teams. The headline number (27%) hides significant variance, and without deeper benchmarks, companies risk applying generic fixes instead of targeted optimization.
This blog will do two things:
- Summarize what Flexera’s 2025 report tells us about cloud waste
- Explore what it doesn’t tell us—the blind spots that matter if you want to benchmark effectively and truly reduce waste
By the end, you’ll have a framework for going beyond Flexera’s high-level averages into actionable, granular cloud cost optimization.
Flexera 2025 on cloud waste: The baselines
Flexera surveyed 759 global cloud decision-makers across industries and company sizes. The findings reflect overall trends in adoption, optimization, and governance.
Key waste figures:
- 27% of IaaS/PaaS spend wasted: Down slightly from prior years, but still massive.
- 24% of SaaS spend wasted: Often from unused licenses or shelfware.
- Cloud cost optimization ranked #1 priority for the ninth year in a row.
Top waste drivers include:
- Idle resources — e.g., dev/test workloads running outside office hours.
- Overprovisioning — allocating more CPU/memory than needed.
- Orphaned resources — unattached volumes, unused IPs, forgotten snapshots.
- Unoptimized licensing — especially with SaaS and enterprise software.
These numbers give organizations a benchmark: if you’re wasting more than ~27%, you’re behind the curve; if you’re wasting less, you’re ahead.
But here’s the catch: averages hide the details that matter most.
What Flexera doesn’t tell you: The blind spots
While Flexera’s report is helpful, it leaves critical gaps for practitioners trying to benchmark cloud waste.
- Variance by industry
A financial services firm with strict compliance workloads wastes differently than a gaming company with bursty usage. Benchmarks need vertical context. - Variance by workload type
Non-production environments (dev, test, staging) can waste 40–60% if unmanaged, while production workloads may waste only 10–15%. Flexera’s averages blur these differences. - Time-based waste patterns
Flexera’s annual snapshot ignores diurnal, weekly, or seasonal waste. For example: retail workloads idle in Q1, peak in Q4; SaaS dev workloads idle on weekends. - Cloud provider differences
AWS, Azure, GCP, and OCI all have unique pricing quirks, discount models, and waste behaviors. A flat 27% benchmark misses provider-specific nuances. - Interdependencies
Shutting down one service may break another (e.g., ETL jobs feeding dashboards). Waste benchmarks must factor in dependency-aware orchestration, not just isolated resources.
Without these nuances, companies may optimize the wrong workloads—or miss the biggest opportunities entirely.
Building a better benchmark framework
To go beyond Flexera’s averages, organizations should track waste with granularity and context. Here’s how:
- Segment by environment
Separate waste benchmarks for dev/test/staging vs. production. - Segment by business pattern
Align benchmarks with real-world usage patterns (compliance cycles, retail seasons, daily/weekly rhythms). - Segment by provider and region
Normalize for AWS vs. Azure vs. GCP differences; include regional pricing variations. - Track interdependencies
Use orchestration to manage dependencies across analytics, ETL, and ERP workloads. - Automate optimization
Manual shutdowns don’t scale. Benchmarks should assume automation for accuracy and sustainability.
This framework shifts benchmarking from static “27% wasted” numbers to dynamic, context-rich insights that guide action.
Conclusion: From averages to action
Flexera’s 2025 report is a valuable compass for understanding cloud waste, but it’s not the map. The average waste numbers (27% IaaS/PaaS, 24% SaaS) are starting points—not endpoints.
Real benchmarking requires:
- Industry-specific context
- Workload-level detail
- Temporal usage patterns
- Provider/regional nuances
- Dependency-aware orchestration
By moving beyond averages, organizations can stop treating cloud waste as an inevitable tax and start treating it as a solvable problem.
The winners in the cloud economy won’t just be the ones who adopt cloud fastest—they’ll be the ones who spend smart, waste less, and continuously optimize with intelligence.
